I'm fascinated by the psychological and creative theory behind interactive narrative. It scares, confuses and enthrals me. Much as there's more and more writing on such topics, I still don't feel like we always understand just what we're doing to the player when we give him a decision to make. I still hold that the logical solution is to tend towards more procedural (albeit less narratively detailed) content: the more dynamic a game world is the more decisions within that world function like real ones (ie a question of knowledge, odds and cause and effect); however more often than not I find myself working with more scripted experiences which send the player off to either (a) or (b).
Where this approach often seems most indelicate is in games with multiple endings. Take Singularity. Without spoiling anything significant, the game's entirely linear, encouraging the player to let the fiction take over: these are the good guys, these are the bad guys, go fight for what's right. At the end, though, you're given a choice: side with the good guys, or the bad guys. For me it renders the rest of the experience completely disingenuous: if this was always an option, why have I only just heard about it? If you wanted me to actually engage my brain and consider the ethical situation I'm presented with, why force me to fight for the good guys the whole way through?
I'm sure we can guess why this decision point is in here. In fact the presentation of the endings tells the tale for us: the good ending is a lavishly presented cut scene; the other two just have naff voice overs. A last minute decision that sounded good on paper, then: of course the player wants to choose who to side with! Of course branching endings are better than straight ones!
It's understandable, because surely more player choice can't be a bad thing? I'm going to argue that, in fact, as far as endings go it can be just that.
Endings in Games vs Endings Everywhere Else
I was considering a scenario for an unannounced project the other day. There was an obvious opportunity for the player to make a decision at the end of the game (let's say he's got a bad guy at gunpoint). It's a secondary character who we can kill off or let live, and we can let that decision affect one or two minor story elements later on. But here's the thing: what decision are we asking the player to make here? Is it between revenge and mercy? Is it between good and evil? Is it between practicality and empathy?
The truth is that in real life (which is what we tend to be modelling with decisions in games) it's all of these things, but the only thing that defines the outcome of the decision once made is raw cause and effect. The more we know about the situation the more able we are to predict its outcome; and when it swings unpredictably (let's say I kill the baddie to be on the safe side, but his brother comes and takes revenge) we accept it as either an oversight on our part, or an unfortunate roll of the dice. This is exactly how such decisions play out in an entirely procedural environment like Spelunky.
But how does this sort of decision point work in the game? Well, for one thing we're going to pre-define the outcome based on what we consider realistic or, perhaps more likely, on what the story and gameplay demand. We're also incapable of understanding what your reasons were for that decision (unless it's a very particular one), which largely negates any useful story consequences we can impose.
Let's take a further example from Jurassic Park: The Game. Major plot spoilers here, but honestly the story's not up to much anyway so don't fret. At the end of the game one of your characters has a decision to make (just like Singularity, it's the only one in the game): save the valuable dino DNA, or save the little girl. I accidentally selected 'Save Little Girl', and I got the happily ever after ending where everyone escapes together and - incredibly - they also happen across a great big bag of gold. Now, the way this is presented suggests going after the DNA was a valid alternative, so I went back and tried again. Turns out when you go after the DNA the character in question winds up getting eaten, and the little girl escapes anyway.
It's obvious enough what Telltale are doing here. We're not being asked to make a practical decision (since both options seem equally risky), we're being asked to make a moral one. The message: be good and your rewards will come; be an arsehole and you'll get digested by a T-Rex. But what if I went after the DNA because I thought the girl was already done for? What if I just put higher ethical value in scientific progress as a whole than in individual human life? What if I panicked and hit the wrong button? Because it's not clear this is a soley ethical dilemma we're confusing this decision; with so many unpredictable factors (the author's hand, the artificial cause and effect of gameplay, the implied freedom of choice) the decision winds up doing very little.
Games vs Films
I don't think there's been a single game with branching endings (since, at least, the internet age) where I didn't immediately go online and see what I missed out on, and I wonder whether that's telling of how unsuccessful games often are in selling to the player that the ending they got is their ending; that it's appropriate to their story. But it also raises some interesting comparisons with film that might shed some light. The entirety of film is 'tainted' by the author's hand. Sure, things usually roll out with a semblance of plausible cause and effect, but unlike games there's nothing in here that's down to chance. There is one route through this narrative, the intended route, and when a character reaches a decision point it's usually made abundantly clear what the alternative was, and why he didn't choose it. The question of 'What if?' is either answered implicitly, or unimportant.
When I dig out the alternate ending to Singularity or Jurassic Park it certainly decreases my identification with the ending I naturally received. The illusion that I had much ownership of this tale is shattered, and often there's little that occurs that isn't predictable (and yet, of course, were it less so we'd be complaining that it wasn't appropriate for a bunch of other reasons). These endings tend to be so run of the mill: we're making a cyber punk shooter, there are three factions in the game, of course you get to choose which one you like at the end!
However, viewing what might have happened can also be illuminating, and it's an option unique to our medium. The good ending in Jurassic Park is an awful lot more meaningful (though sadly no more interesting) once I know what the alternative was. When we consider the difference between the illusion of freedom and true freedom, knowing what your other options were at least confirms that you had freedom of a meaningful sort.
So back on topic, what did I do in my execution scenario? Well, I thought about playing the odds. Make it clear to the player what possible ramifications exist (ie the vengeful brother), and then assign it a probability. For 75% of players it's a risk worth taking; the rest get unlucky. In the end I decided there were too many factors at work. I'd rather provide a linear experience that's at least true to itself (whose drama plays out in a meaningful way) than provide the player a decision point just because I can, and risk him being alienated because I haven't properly accounted for his motivation.
There's a danger, in this piece, that I'm just complaining about all the options available to a narrative designer, without proposing solutions. Given my theoretical leaning towards procedural play there's no doubt some truth to that, but I'm not happy leaving it there.
At Southbank, I always push my story design students to ensure any branching they consider is appropriate to the player. There's no point giving him ownership of the story if it doesn't branch in a way that's more meaningful for him than it might be for differently disposed players, and that principle carries over to this discussion. Singularity: there's nothing wrong with a tightly crafted linear experience, so just bite the bullet.
The game I'm working on does have multiple endings. It's a risk. But we're conscious of that fact, and we're working hard from the get go to ensure it's based on meaningful player decisions. Perhaps more importantly we're not throwing in decisions just for the sake of it. When a player comes into our game I want him to know what his options are, and what they're not; and while I don't want him to see the ending coming, I want it to feel appropriate nonetheless.
We'll find out whether we succeed early next year.